328 BOTANICAL GAZETTE [APRIL 
proportion of the two kinds of seeds. Nearly as large increases 
were obtained where the number of crossed seeds greatly exceeded 
the selfed. These data obtained from reciprocal crosses fully 
substantiate the results reported by Cortins and Kempton, and 
altogether show that CORRENS 
(5) was not wholly correct in 
stating that crossing does not 
immediately alter the size of 
seeds in maize. 
So far as I know, maize is 
the only plant in which this 
manifestation of heterosis has 
been demonstrated. Since the 
main facts of xenia and heterosis 
as determined in maize do not 
differ essentially from the results 
obtained in other plants, there 
is every reason to suppose that 
increased endosperm develop- 
ment resulting from crossing is 
a phenomenon which may 
occur in many, if not all, 
other angiosperms where 
double fertilization takes place. 
Granted that this is so, what 
bearing do these facts have 
upon the puzzling problem of 
double fertilization in endo- 
sperm formation ? 
Nemec (13) has suggested, 
Fic. 2.—Machine used for weighing oo of accounting for = 
seeds. origin of the process of endo- 
sperm hybridization, that it isan 
adaptation resulting in an alteration of the food supply to accord 
with the properties of a hybrid embryo. His own statements in 
regard to the matter are as follows: 
ae cjpaite of hybridization, the embryo and the endosperm are assured the 
same physiological properties only when the endosperm fusion nucleus as well 
